Geoarchaeology and Late Quaternary Geomorphology of the Middle South Platte River, Northeastern Colorado

Vance T. Holliday Department of Geography, Science Hall, University of Wisconsin, Madison, WI 53706

Investigations were conducted along the middle South Platte River to better define the geomorphic contexts of Paleoindian sites and to reconstruct the alluvial and eolian geochronology. Paleoindian sites are associated with the Kersey terrace (the downstream equivalent of the Broadway terrace). The Kersey alluvium was deposited during Clovis occupation and the surface stabilized by 10,000 B.P. Post-Clovis sites post-date aggradation and stream downcutting may have started as early as 10,500 B.P. Subsequent floodplain development and downcutting formed the Kuner terrace (thepossible downstream equivalent of the Piney Creek terrace) no later than 3000 B.P. and the Hardin terrace probably within the last 1000 years. Soils on the Kersey terrace are Ustochrepts (gravelly alluvium) or Haplustalfs (sandy and clayey alluvium). Soils on the Kuner terrace are cumulic Ustorthents and Ustochrepts. Soils on the Hardin terrace are Ustorthents with no obvious horizonation. Eolian sands began accumulating in the region by 10,000 B.P., but most are probably late Holocene deposits and are indicative of drier post-Pleistocene climate. Correlations with deposits in low order tributaries and other drainages can be difficult to make a) using soils because soil development varies as a function of parent material texture and b) because aggradation and degradation may be out-of-phase.

INTRODUCTION necessary for establishing the late-Quaternary geochronology of the middle South Platte. The Great Plains of eastern Colorado con- tain many well known Paleoindian sites (e.g., Dick and Mountain, 1960; Wheat, 1972,1979; PREVIOUS STUDIES Wilmsen and Roberts, 1978; Stanford, 1978). Some are along the middle South Platte River, Paleoindian sites in the study area have in Weld County, including the Dent site been investigated for over 50 years. The Dent (Wormington, 1957), the Jurgens and Frazier site, near Milliken (Figure 11, was found in sites (Wheat, 19791, and the Powars site (Rob- 1932. Bilgery (n.d.1, followed by Figgins erts, 1937).Much archaeological research was (1933), conducted the first excavations. They carried out at some of these sites, but little is found the remains of numerous mammoths known about their geomorphic settings and associated with stone tools, including three geologic histories. Clovis points. The site was exposed in an ar- This paper presents the results of some geo- royo incised into a narrow terrace of the South archaeological field work at and in the area of Platte located between an outcrop of Creta- the Paleoindian sites in Weld County. This ceous sandstone and the modern floodplain of research was carried out for several reasons. the river (Figures 1 and 2). Few details are Geoarchaeological data are important for in- available concerning the archaeology and the terpreting the occupation and post-occupation geology of the site is unclear because of both histories of the sites and for understanding natural and artificial disturbance. Most im- the regional setting during the Paleoindian portantly, it is not certain whether the bone occupation of the area. In addition, an under- was deposited in an arroyo cut into the terrace standing of the late-Quaternary geomorphic or whether the bone was deposited within sed- history of this portion of the South Platte iments comprising the terrace and exposed could aid future archaeological research and only by younger arroyo cutting. Malde (19541, complement geomorphic research in other Spikard (19721, and Haynes (1974)conducted reaches of the South Platte. Finally, the ar- limited additional research. Cassells (1983: chaeological information provides the data 44-49) has synthesized and reevaluated

Geoarchaeology: An International Journal, Vol. 2, No. 4, 317-329 (1987) 01987 by John Wiley & Sons, Inc. CCC 0883-6353/86/040317-13$04.00 MIDDLE SOUTH PLATTE RIVER

Figure 1. Map of the middle South Platte River in the Greeley, Colorado area showing the general surficial geology and location of the Kersey area, the Dent site, and line-of-section A-A' through Dent (Figure 2). Inset shows Colorado with the principal physiographic features mentioned in the text and locations of Denver (D) and Greeley (G). much of the existing data. Some consider Dent The site consisted of three activity areas, in- to be the first discovery of fluted projectile cluding a butchering or processing area con- points associated with mammoth remains in taining the disarticulated remains of extinct North America (Wormington, 1957), but Cas- bison, a short-term camping area, and a long- sells (1983) presents evidence that the bone is term camp or habitation site. The three areas redeposited (discussed below). are linked by the presence of Kersey projectile The Jurgens and Frazier sites, both discov- points in each (Wheat, 1979). ered by Frank Frazier, are later Paleoindian Two other important Paleoindian sites oc- sites near Kersey (Wheat, 1979; Wormington, cur near Kersey, but there is little informa- 1984) (Figure 3), where they occur just below tion on them. The Powars site, found in dunes the surface of a terrace of the South Platte, south of the river (Figures 2 and 3) and exca- about 1.5 km apart (Figures 2 and 3). The vated in the 1930s, was a campsite which Frazier site, briefly described by Wormington yielded Folsom artifacts (Roberts, 1937). The (1984) and investigated by Malde (1984),con- Klein site, located just north of Powars and on tained the disarticulated and butchered re- the same terrace as Jurgens and Frazier (Fig- mains of extinct bison (Bison antiquus) and ures 2 and 3), consisted of surface finds of associated lithic artifacts, including Agate Clovis projectile points exposed by deep plow- Basin projectile points. The Jurgens site was ing (Adrienne Anderson, William Klein, excavated by Wheat (1979) and Hall (n.d.1 Frank Frazier, personal communication, carried out some geoarchaeological research. 1978). Limited trenching yielded no cultural

31 8 VOL. 2, NO. 4 MIDDLE SOUTH PLATTE RIVER

A A' Qf =floodplain alluvium elev (m) Qh=Hardin alluvium 1460-- SQ'/Dent site Qe=eolian seds. Qku = Kuner a1 luvium Qke= Kersey alluvium QI= Louviers alluvium - --=bedrock 4w €*

Jurgens €31 Frazier Klein Powars B 1410

1370- -0 0 3000ft vert exog. = 251 Figure 2. Geologic cross sections through the Dent site (A-A'; Figure 1)(above) and across the South Platte at Kersey (B-B'; Figure 3) (below) with the relative locations of the Paleoindian sites in the Kersey area. Data on the depth to bedrock at Kersey are from Smith et al. (1964:Plate 10, section J-J').

UQf f loodplain alluvium

OQh Hardin alluvium nQeeolian deDosits

OQkeKersey alluvium - terrace scarp

...I. gravel ridge _- _- paleochannel - archaeological - J site soil profile 7. locality

0 2 km

-4 -4

Figure 3. Map of the surficial geology in the Kersey area with the locations of Paleoindian sites, soil profile localities, and line-of-section B -B'.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 31 9 MIDDLE SOUTH PLATE RIVER material (Adrienne Anderson, Vance Haynes, of pediments, strath terraces, and fill terraces personal communication, 1984).Clovis projec- along with eolian sediments in most areas tile points were also recovered from gravel (e.g., Scott, 1963,1965;Machette, 1975a; Van pits in the same terrace in the Kersey area Horn, 1976; Colton, 1978). Hunt (1954) and (Wheat, 1979; Frank Frazier, personal com- Scott (1963), working on the South Platte in munication, 1978). and upstream from Denver, and Bryan and Bryan and Ray (1940), who were concerned Ray (1940), working on the Cache la Poudre primarily with the geologic history of the fa- system and middle South Platte, established mous Lindenmeier archaeological site, north the classic Quaternary stratigraphic frame- of Ft. Collins, have also provided regional geo- work for the Piedmont. archaeologic and late-Quaternary geomor- phic data for Paleoindian sites in the area. METHODS The geologic and geomorphic features along the South Platte in the Kersey area were con- Most field work was conducted in the Mil- sidered in their work and they established the liken and Kersey areas, but localities between first model for the geologic evolution of the these two areas were also studied, including middle South Platte River and proposed ter- the dune fields that border the valley. Numer- minology for many of the geomorphic surfaces ous natural and artificial exposures were ex- of the region. Their terminology is still ap- amined (Figure 3) and all were described us- plied in the region and was employed in this ing standard pedologic terminology (Soil study. Survey Staff, 1975; Guthrie and Witty, 1982; Birkeland, 1984) (Tables I-IV). The soils are classified to the Great Group level following SETTING Soil Taxonomy (Soil Survey Staff, 1975) based The study area is within the Colorado Pied- on field criteria. There is no laboratory data mont section of the Great Plains physio- available because initially the scope of the graphic province (Figure 1) (Fenneman, study was so small as to preclude laboratory 1931). The Colorado Piedmont is both a topo- analyses. As the investigations expanded it graphic and structural basin that separates became apparent that laboratory data would the High Plains to the east from the Rocky contribute no significant information beyond Mountains to the west (Figure 1). The Pied- further characterization of the soils and mont was once covered by Cenozoic sediments sediments. which presently underlie the High Plains. Erosion by the South Platte and Arkansas Rivers and tributaries throughout the late LATE QUATERNARY STRATIGRAPHY AND GEOMORPHOLOGY Cenozoic removed most of the Cenozoic cover and cut into Mesozoic deposits, thereby form- The most obvious and best preserved geo- ing the Piedmont. Cretaceous shales, sand- logic evidence of the late Quaternary history stones, limestones, and coals and remnants of of the South Platte valley in the study area some Paleocene deposits comprise the bedrock occurs on the south side of the river in the in the study area (Hunt, 1974; Colton, 1978). Kersey area (Figure 3). This is fortunate be- The South Platte heads in the Rocky Moun- cause data from the archaeological sites men- tains, flowing north and east across the Pied- tioned above can be used for estimating the mont and High Plains (Figure 1). Several ma- ages of some of the landforms and deposits. jor tributaries within the Piedmont and study Three terraces occur in this area and exten- area include the Big Thompson, joining west sive eolian deposits cover the uplands imme- of Milliken, and the Cache la Poudre, entering diately south of Kersey and north of the river west of Greeley (Figure 1). (Figures 2 and 3). Bryan and Ray (1940) refer Episodic degradation and aggradation to the highest terrace as the Kersey and the throughout the Quaternary has left a sequence next lower terrace as the Kuner (after a small

320 VOL. 2, NO. 4 MIDDLE SOUTH PLATTE RIVER

town downstream). Frazier (personal commu- illustrate over 50 m of valley fill in the South nication, 1978) named the lowest terrace the Platte at Kersey (Figure 21, but they do not Hardin (after another small town down- indicate the thickness of the Kersey compo- stream). nent. In all exposures observed in this study, In other parts of the study area the terraces the Kersey alluvium is at least 10 m thick. were more difficult to identify because eolian On the north side of the South Platte and sediments north of the river bury the terraces Cache la Poudre Rivers, the highest terrace is (Figure 2). Only portions of the highest ter- as much as 6 m higher than the Kersey ter- race could confidently be identified. Else- race on the south side of the South Platte. where (e.g., south of Greeley), some of the Because of the height of this terrace, Bryan terraces are not preserved or construction ac- and Ray (194056) named this the Pleasant tivity has obscured or destroyed the geomor- Valley terrace and considered it to be older phic relationships. At Dent all attempts at than the Kersey. Scott (1963:32), however, examining the soil formed in the Kersey ter- suggests that the Pleasant Valley and Kersey race (into which the bone-bearing arroyo was deposits are coeval based on his observations cut) were frustrated by heavy disturbance and those of Hunt (1954:109). Scott speculates from railroad construction. that alluvium was brought into the South Platte by the Cache la Poudre faster than the South Platte could carry the material, resulting in a Kersey Terrace “piling up” effect along the Cache la Poudre The Kersey terrace is the oldest and most just above and below its mouth. Surface pro- obvious late Quaternary geologic feature in files of the two terraces show that they con- the area. The deposits consist of bedded sand verge downstream (Figure 41, suggesting that and gravel and the primary clast constituents Scott (1963) may be correct. are granite, quartz, and pegmatite. The Ker- Several significant variations in the lithol- sey appears to be a fill terrace and is generally ogy and geomorphology of the Kersey terrace broad and nearly level. The surface is approxi- occur on the south side of the South Platte at mately 10 m above the South Platte floodplain Kersey. The sediments of the northern por- and 2.5 to 3.5 km wide near Kersey and as tion of the terrace are noticeably coarser, great as 4 km wide on the south to southeast where medium to coarse gravel is common. To side of the South Platte from the Greeley to the south and extending under the dunes the the Milliken area (Figures 1and 3). However, sediments are generally sandy with some fine the depth of the Kersey alluvium is unknown. to medium gravel. Smith et al. (1964: Plate 10, section J-J’) The surface of the northern portion of the

4w E* Greeley Kersey 4 4 e I ev.(m )

-1420

-1410 -

conf I ence 2 km u O.. - O.. cache 10 P oudre o imi --- Figure 4. Stream terrace profiles along the South Platte and Cache la Poudre Rivers showing the merging of the “Pleasant Valley” terrace and Kersey terrace.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 321 MIDDLE SOUTH PLATTE RIVER terrace has ridge-and-swale topography. bly developed in silty sediments and contains Wheat (1979; personal communication, 1978) an A-Bt-Bk-Ck profile (Table I, 5cm+). shows that both the Jurgens and Frazier sites The variation in soil development across are situated along ridges and swales. On ae- the Kersey terrace is probably due to the par- rial photographs numerous gravel ridges, ent material variations across this surface. truncated during land-leveling, are exposed The stronger soil development, as expressed and are especially common near the valley by B horizon development and content of illu- center (Figure 3). Relief between the ridges vial clay, occurs in the finer-grained parent and swales is unknown but the maximum re- material. Machette (1975a) noted similar pe- lief was no more than 3 m (Wheat, 1979). dologic variations upstream from the Kersey These features probably formed when a area. In addition, the mantle of eolian sedi- braided stream existed on the Kersey surface ment at localities 2 and 3 could have contrib- after aggradation of the fill. Exposed ridges uted clay and carbonate to the underlying soil. are less common on the southern portion of The archeological sites on the Kersey sur- the surface. Rather, a broad swale up to 1km face provide valuable information for dating wide and trending east -west is apparent on terrace formation. The Clovis artifacts found topographic maps (Wheat, 1979) (Figures 2 within the Kersey alluvium at the Klein site and 3). This variation in channel morphology and in gravel pits indicate that aggradation may explain the lithologic changes noted was still in progress during the Clovis occupa- across the terrace, with the finer sediments tion, which dates between 11,500 and 11,000 deposited by the wide meandering stream and B.P. at other sites on the Great Plains (Frison, the coarser material deposited by the braided 1978; Haynes, 1980).The Frazier and Powars stream. No evidence was found to indicate sites are probably about the same age and when or in what sequence this change in postdate aggradation of the Kersey alluvium. channel pattern occurred. The Frazier site yielded two radiocarbon ages Crabb (1980) maps the dominant soil series on soil organic matter, both 9600 2 130 B.P. on the Kersey terrace as the Altvan and As- (SMU-31,32; NaOH-soluble fraction) (Haynes calon loams and Nunn and Cacono clay loams. and Haas, 1974; Wheat, 1979), and projectile All of these soils are Argiustolls, character- points of the Agate Basin variety. The radio- ized by A-Bt-Bk-C profiles to a depth of about carbon ages are considered minimum dates 1.5 m. However, such well developed soils for the Agate Basin horizon (Haynes and were not observed on the Kersey terrace in the Haas, 1974). Wormington (1984) notes that Kersey area. Soils are significantly less well the radiocarbon samples came from an organic- developed near the valley center compared to rich soil horizon. This horizon is apparently at those along the valley margin. At locality 1, or above the level of the occupation, but the just east-southeast of the Jurgens site (Figure nature of the horizon and the exact relation- 3), historic unconsolidated material overlies a ship of the samples to the stratigraphy and truncated, buried Ustochrept that has a Bw-C archaeology is unclear. Frison (1978) and profile (Table I, 18cm+) developed in grav- Frison and Stanford (1982) date the Agate elly, sandy alluvium. In the swales at Frazier Basin culture to over 10,000 B.P., probably and Jurgens the soils formed in clayey allu- within the period 10,500to 10,000 B.P. Folsom vium and are characterized by gleying (Wheat, material was recovered at the Powars site 1979; Malde, 1984). At locality 2, east-south- from dunes overlying the Kersey terrace. The east of Kersey (Figure 2) another truncated Folsom occupation at other sites on the Great though more strongly developed buried soil Plains has about the same age range as Agate was observed. This soil formed in sand and is Basin and perhaps extending back to 11,000 probably a Haplustalf with a Bt-Btk-C profile B.P. (Frison, 1978; Frison and Stanford, (Table I, 82cm+). North of the Cache la 1982). These data indicate that the Kersey Poudre at locality 3 (Figure 3) is another bur- surface was available for occupation no later ied though complete Haplustalf which proba- than about 10,000 B.P. and probably earlier.

322 VOL. 2, NO.4 MIDDLE SOUTH PLATE RIVER

Table I. DescriDtions* of soils of the Kersev terrace. Depth, cm Horizon DryColor Texture Structure Eff Bndy Comments Locality 1

(see Figure 3 for location)

0-18 C lOYR413 Lm-fS gr non as eolian cover? 18-28 2Bwb lOYR312 gr,m-fSL wSBK non cw to SCL 28-50 2Clb lOYR 313 gr,m-fSL vwSBK non cs 50-75+ 2C2b lOYR 313 m-fLS to vwSBK non SL

Locality 2

(see Figure 3 for location)

0-5 C lOYR 514 fS non cs eolian sand 5-20 Abl lOYR 413 m-fS non cs eolian sand 20 - 42 Cbl lOYR 414 C-fS non cs eolian sand 42 - 56 Ab2 lOYR 413 Lm-B non cs eolian sand 56 - 67 Clb2 lOYR 414 m-fS non cs eolian sand 67-80 C2b2 lOYR 415 m-fS non aw eolian sand 80-82 C3b2 7.5YR414 Lm-B non aw eolian sand 82- 95 2Btb3 7.5YR3/2 mSCL non cw 95-110 2Btkb3 7.5YR 312 mSCL es cw Stage I1 carb; noncalc ped interiors 110-140 3Bkb3 lOYR812 fSCL M ev ds Stage I11 carb; noncalc ped interiors 140-190-1 3c lOYR 512 fS sg non Locality 3

(see Figure 3 for location)

0-5 C lOYR 413 SiCL gr non cs modern flood deposits? 5- 13 Abl lOYR312 CL P1 non cs 13- 19 Btlb lOYR 3/3 Sic wPr-mSBK non cw 19-34 Bt2b lOYR 313 Sic wPr-sSBK non aw 34- 45 Bklb lOYR 6/2 Sic M ev cw Stage I11 carb; noncalc ped interiors 45-90 Bk2b 10YR 7/2 Sic M ev aw Stage I11 carb; noncalc ped interiors 90 - 105 Ckb lOYR 614 B M ew aw Stage I carb; (bedded) noncalc ped interiors: 105- 125 Cb lOYR 6f4 fS M non aw (bedded) *Abbreviations for descriptions: Texture: S = sand f = fine m = medium c = coarse LS = loamy sand SL = sandy loam SCL = sandy clay loam L = loam CL = clay loam Sic = silty clay SiCL = silty clay loam, gr = gravelly Structure: Grade, v = very w = weak m = moderate s = strong Type, SBK = subangular blocky Pr = prismatic P1 = platy M = massive sg = single grain gr = granular Eff (effervescence from reaction w/dilute HCl): non = noncalcareous ew = weakly effervescent es = strongly effervescent ev = violently effervescent Bndy (boundary): distinctness, a = abrupt c = clear g = gradual d = diffuse topography, s = smooth w = wavy

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 323 MIDDLE SOUTH PLATTE RIVER

There is little evidence to indicate when the Kuner Terrace South Platte began to incise and isolate Ker- The Kuner terrace is another fill terrace, sey alluvium as a terrace. At the Frazier site Malde (1984) notes that the Agate Basin occu- situated about 5 to 6 m above the modern pation is at the base of eolian sand immedi- floodplain of the South Platte. In the Kersey ately above a gleyed sandy clay loam. The area the Kuner terrace is represented only by gleying apparently occurred when the water remnants at the foot of the Kersey scarp and table was just below the surface, i.e., prior to the best preserved remnant is north of Ker- downcutting. Malde (1984) believes that the sey, just below the Jurgens site (Figures 2 eolian sedimentation could not take place and 3). The terrace remnants in this area are so narrow that on soil maps they are included until downcutting began and, therefore, with the Kersey terrace. The Kuner alluvium downcutting began prior to the occupation of the Frazier site. There is virtually no data, is similar to the Kersey in lithology and tex- ture. The thickness of the fill is unknown. however, on the precise relationship of the The Kuner terrace exhibits ridge and swale archaeology to the sediments at Frazier. Downcutting preceded formation of the Kuner topography similar to that observed on the terrace, but otherwise age estimates for Kersey terrace and soil development varies considerably depending on the position of the downcutting must remain speculative. soil on the ridges and swales. The best devel- The Jurgens site is the youngest of the oped soil on a ridge is an Ustochrept, at local- Paleoindian sites in the area, with a radiocar- bon age of 9070 * 90 B.P. (SI-3726) (Wheat, ity 4 (Figure 3), with an A-Bw-Bk-C profile 1979). The occupation zone was apparently mantled by eolian sand (Table 11,27 cm+).In above a gleyed zone similar to that observed the nearby swale the soil is a cumulic Ustor- at Frazier, but otherwise the site yielded little thent, composed of a dark brown, organic- data bearing on the history of the river. rich, fine sandy clay up to 2 m thick, overlying a brown loamy sand. The soil on the Kuner As noted, Cassells (1983) presents evidence terrace Dent is an Ustochrept with an Ap- suggesting that the Dent material was rede- at posited. In the Dent area the next higher ter- Bw-C profile formed in fine sands (Table 11). race above the Kersey (on top of the bedrock On the Kuner terrace, therefore, the best de- immediately west of the site) is the Louviers veloped soils occur on higher ridges with eo- (Figure 2). The Louviers sediments differ lian sand caps. In the swales the soils are from the Kersey deposits in that the former cumulic with little horizonation. are much coarser, typically containing coarse The Kuner terrace was present by the mid- gravel, and the undersides of the gravel typi- dle Holocene. No radiocarbon ages are avail- cally have coats of calcium carbonate. These able, but a few diagnostic artifacts belonging characteristics of the Louviers are also noted to the Duncan and McKean culture complexes by Machette (197513) for areas upstream. At were recovered from the surface (Frank Dent, coarse gravel with randomly oriented Frazier, personal communication, 1978). carbonate was found among the mammoth These artifacts are middle Archaic and date bones. This led Frazier (cited in Cassells, between 5000 and 3000 B.P. (Frison, 1978). 1983) to believe that the bone was washed in This suggests that the terrace was created by along with Louviers gravel. This interpreta- at least 3000 B.P. tion is supported if the arroyo that yielded the finds initially developed before the bone accu- Hardin Terrace mulated. This arroyo, graded to a lower level than the Kersey terrace, developed after the The Hardin terrace is usually 1to 2 m above South Platte started downcutting, which oc- the floodplain of the South Platte and quite curred after the Clovis occupation. On the narrow (Figures 2 and 3). It is often obscured other hand, if the bone and gravel were part of by irrigation ditches. The Hardin is another the Kersey alluvium, then it was probably fill terrace similar in lithology to the Kersey redeposited during the Clovis occupation. and Kuner, although generally finer in tex-

324 VOL. 2, NO. 4 MIDDLE SOUTH PLATTE RIVER

Table 11. DescriDtions of soils of the Kuner terrace. Refer to Table I for abbreviations.

~~~ Depth, cm Horizon Dry Color Texture Structure Eff Bndy Comments Locality 4 (see Figure 3 for location)

0-8 C lOYR513 fSL sg non as modern slopewash 8-12 Abl lOYR313 m-fSL vwSBK non ds modern slopewash 12-27 Cbl lOYR313 m-fLS sg non cs modern 27-35 2Ab2 lOYR312 fSCLtoL vwPr-wSBK non cw slopewash 35-45 2Bwb2 lOYR313 L toCL wPr-wSBK non aw 45 - 75 2Bkb2 lOYR514 fSCL M es aw Stage I carb; noncalc ped interiors 75-90t 3Cb2 1OYR 516 s,c-fS sg non

Dent Site

(see Figure 1 for location)

0-35 Ap lOYR 313 m-fS sg non cw mottled wl 30% 314 35 - 50 CBw lOYR414 LfS sg non cs 50-70 C 1OYR 4.514 m-.c-mS SB non

ture with fewer and smaller gravels. The soils The eolian deposits in the Kersey area are in the Hardin alluvium at locality 5 (Figure 3) considered to be of Holocene age because they and Dent (Table 111) are probably Ustiflu- overlie the soil developed on the latest Pleis- vents and exhibit no obvious horizonation. tocene Kersey terrace. The presence of Folsom The very minimally developed soils on the material in dune sediments at the Powars site Hardin terrace suggest it is quite young and, indicates that some eolian sediment began to based on comparisons with the soils of the accumulate by the beginning of the Holocene. Kuner terrace, probably less than 1000 years This is also demonstrated by the occurrence of old. Agate Basin material at the base of eolian sediments in a swale at the Frazier site (Malde, 1984). However, significant accumu- Eolian Deposits lations of eolian sand and development of the The eolian deposits in the study area consist dune field probably did not occur until the late of extensive sand dunes to the south of Kersey Holocene because soil development within and dunes and probably loess to the north and at the top of the sands at localities 2 and 6 (Figures 1, 2, and 3). The sands mantle the (Figure 3) is very weak (Tables I and IV). Kersey terrace on both sides of the river (Fig- Muhs (1985), in a study of dunes in the region ure Z), but neither the extent of the Kersey and eastward into Nebraska, also considered terrace under this material nor the depth of the dunes to be late Holocene features, and the eolian deposits could be determined. mostly no more than 3000 years old. This Smith et al. (1964:25) considered the deposits would allow at least 7000 years for soil quite shallow and ignored them and mapped development on the Kersey terrace where it only the underlying Cretaceous sandstone. is buried by dunes. The Holocene eolian sedi- No outcrops of bedrock were located in the ments in the study area are significant be- project area, and adequate exposures of the cause they are indicative of a climate drier eolian deposits themselves were difficult to than that of the late Pleistocene. locate. Muhs (1985) also noted a distinct, though

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 325 MIDDLE SOUTH PLATTE RIVER

Table 111. Descriptions of soils of the Hardin terrace. Refer to Table I for abbreviations. Depth. cm Horizon Dry Color Texture Structure Eff Bndy Locality 5 (see Figure 3 for location)

0-13 c1 7.5YR 514 f-mLS sg non aw 13 - 16 c2 7.5YR 312 Lc-fS sg non aw 16-38 c3 7.5YR 313 Lf-cs sg non aw 38-51 C4 7.5YR314 m,Lc-fS sg non aw 51 -62 c5 7.5YR 312 p,c-fSL sg non aw 62-90+ C6 7.5YR 312 c-fsL sg non aw Dent Site (see Figure 1 for location)

0-40 c1 lOYR 414 bedded sg-M non aw 313,312 S&C 40-60+ c2 lOYR 412 fS sg non aw

Table IV. Description of soil in sand dune, locality 6. Refer to Table I for abbreviations. (see Figure 3 for location) Depth. cm Horizon Dry Color Texture Structure Eff Bndy 0-74 AP lOYR 414 C-fS sg non gs 74-125 c1 lOYR 5/4 C-fS sg non gs 125 - 250+ c2 lOYR 6/4 C-fS sg non localized buried soil within the dunes, charac- and Broadway terraces are also generally terized by a weak Bt horizon and separating comparable, although the strongest soil de- earlier Holocene sands from the late Holocene veloped on the Kersey terrace has a Bt horizon sands. This buried soil was not observed in the and the strongest soil on the Broadway ter- Kersey area, but given the presence of some race does not (Scott, 1963; Machette, 1975a,b; early Holocene sands, this soil may well occur Van Horn, 1976).This is probably because the locally and would be an excellent stratigraphic parent material in the Kersey area is finer- marker. grained. The coarse-grained facies of both the Kersey and Broadway deposits (Machette 1975a,b) exhibit weakly developed soils with REGIONAL CORRELATION AND either A-C or A-Bw-C profiles. COMPARISONS Correlating the Kuner and Hardin terraces The data from the Kersey area can be com- is difficult because these surfaces cannot be pared with the late Quaternary stratigraphy traced directly into other areas. The strongest and geoarcheology in neighboring areas. soil developed on the Kuner terrace is similar Scott (1963), Smith et al. (1964), and Colton to the strongest soil developed on the Piney (1978) correlate the Broadway terrace of the Creek terrace in the Denver area, which is Denver area with the Kersey terrace. Clearly, also below the Broadway terrace and thought these terms apply to the same surface because to be of middle Holocene age (Scott, 1963). the terrace can be followed continuously from These two surfaces may therefore be related Kersey to Denver. The Kersey terrace can and were correlated with one another by Hunt also be followed downstream toward Ft. Mor- (1954). Gardner (1967) identified the Kuner gan, where it was studied by Gardner (1967). terrace downstream from the Kersey area, The soils formed in the alluvium of the Kersey but direct correlations between these areas

326 VOL. 2, NO. 4 MIDDLE SOUTH PLATTE RIVER cannot be made and there are no other ade- regions, may well act independently of one quate data with which to compare the down- another in their response to both internal and stream terrace. external factors (Schumm, 1973). The post-Piney Creek deposits of Scott (1963) in the Denver area exhibit very mini- SUMMARY mal soil development, but yielded charcoal which gave a radiocarbon age of about 1500 The middle South Platte River in the Colo- B.P. Machette (1975a,b) identifies a thin A rado Piedmont was the scene of relatively horizon in post-Piney Creek soils northwest of intense occupation by Paleoindian cultures. Denver. The Hardin soils have no horizona- In the Kersey area, Paleoindian sites occur on tion and are therefore probably of the same the Kersey terrace, which is the oldest, high- age or somewhat younger than the post-Piney est, and best preserved of three late Quater- Creek sediments. Gardner (1967) identified nary terraces of the middle South Platte the Hurley terrace below what he considered River. The Kersey terrace is the downstream the Kuner terrace downstream from Kersey. equivalent of the Broadway terrace of the As was the case for the Kuner terrace, there Denver area. The Pleasant Valley terrace of are no data with which to correlate the Hardin Bryan and Ray (19401, though somewhat and the Hurley terraces. higher in elevation than the Kersey, is now On a regional scale Wilmsen and Roberts considered to be the equivalent of the Kersey (1978)correlated deposits at the Lindenmeier on the north side of the South Platte at Kersey site with the South Platte alluvial chronology as well as upstream along the lower Cache la of Scott (1963). At Lindenmeier the alluvium Poudre. overlying the Folsom occupation zone is corre- Aggradation of the Kersey alluvium ended lated with the Broadway-Kersey alluvium. between 11,500 and 10,000 B.P. Clovis However, in the Kersey area, Folsom mate- (11,500-11,000 B.P.) material occurs in the rial at the Powars site is in dune sand resting upper Kersey alluvium at the Klein site and on top of the Kersey terrace. Therefore, both other scattered localities and was often rede- Folsom-age and post-Folsom-age sediments posited, including the material found at the at Lindenmeier are younger than the Broad- Dent site. In its final stages of aggradation, way - Kersey alluvium. This miscorrelation the South Platte had both a wide, meandering illustrates the problems inherent in attempt- channel and a braided channel, although the ing correlations of alluvial deposits between temporal relationships of these two channels sites in small tributaries of small valleys with is unknown. The South Platte may have major drainages on the plains. These prob- started downcutting by 9000 B.P. A Folsom lems are further exaggerated when sites near (11,000- 10,000 B.P.) occupation is present at the mountain front are compared to those far- the Powars site in eolian sands that cover the ther out on the plains. Given the variations in Kersey terrace. Agate Basin (10,500- 10,000 local geomorphic and hydrologic situations in B.P.) and Kersey (9000 B.P.) cultures occu- the small tributaries, depositional and ero- pied the ridge-and-swale surface of the Ker- sional cycles may well out-of-phase and other- sey terrace at the Frazier and Jurgens sites, wise have no relationship with those in the respectively. Soils on the Kersey terrace are larger drainages. Ustochrepts with A-Bw-C profiles and Hap- Eddy et al. (1984) applied the South Platte lustalfs with A-Bt-Bk-C profiles, varying sig- stratigraphic sequence of Scott (1963) to the nificantly as the texture of the parent mate- Arkansas River in southeastern Colorado. rial changes. Stronger soil development oc- The results could be very misleading in mak- curs in the finer grained facies. ing geochronological or environmental inter- Two narrow, discontinuous terraces occur pretations, Like the problems in correlating below the Kersey. Of these, the higher Kuner different parts of a drainage, separate drain- terrace, which sometimes contains Archaic ages, even though in similar environments or cultural material, was created before 3000

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 327 MIDDLE SOUTH PLATTE RIVER

B.P. Ridge-and-swale topography is also ap- their considerable knowledge of the archaeology and parent on this surface. Ustochrepts with geology of the middle South Platte. Adrienne Anderson (National Park Service, Denver) also provided valuable A-Bw-C or A-Bw-Bk-C profiles have formed information and field assistance. William Klein (Ker- on the ridges and cumulic Ustorthents formed sey) and Herbert Bernhardt (Milliken) kindly allowed in the swales. The lower Hardin terrace prob- access to their land. Adrienne Anderson and Peter Birkeland (University of Colorado, Boulder) reviewed ably developed within the last 1000 years. earlier drafts of this paper and Dan Muhs (U.S.G.S., Soils on the Hardin terrace are Ustorthents Denver), Richard Reider (University of Wyoming), and with no significant horizon development. Cor- an anonymous reviewer provided helpful comments on relations of these two terraces with others the final manuscript. described along the South Platte are difficult to make. The Kuner may be the downstream equivalent to the Piney Creek of the Denver REFERENCES area. The Hardin terrace appears to be youn- Bilgery, C. (n.d.). Evidences of Pleistocene Man in the ger than the post-Piney Creek terrace identi- Denver Basin. Unpublished manuscript, University fied near Denver. of Colorado Museum, Boulder. Birkeland, P. W. (1984).Soils and Geomorphology. New Eolian sand is extensive on the uplands on York: Oxford University Press, 372 pp. either side of the river. Some of these sedi- Bryan, K., and Ray, L. (1940). Geologic antiquity of the ments began accumulating by 10,000 B.P., Lindenmeier site in Colorado. Smithsonian Miscella- but most of the sands are probably late Holo- neous Collections 99, 1 - 76. Cassells, E.S. (1983). The Archaeology of Colorado. cene deposits. The sands are indicative of a Boulder: Johnson Publishing, 325 pp. drier Holocene climate compared to that of the Colton, R.B. (1978). Geologic map of the Boulder-Fort late Pleistocene. Collins-Greeley Area, Colorado. U.S.Geological Sur- vey Map I-855-G. Correlations of sediments in the middle Crabb, J.A. (1980). Soil Survey of Weld County, Colo- South Platte with those in low order tributar- rado, Southern Part. US. Department of Agricul- ies or in other drainages must be done with ture, Soil Conservation Service. Dick, H.W., and Mountain, B. (1960).The Claypool site: extreme caution. Soils can be useful tools for a Cody Complex site in northeastern Colorado. making correlations, but the data from the American Antiquity 26, 223-235. middle South Platte show that soil develop- Eddy, F.W., Oberlin, R.E., and Farmer, T.R. (1984). ment on a surface can vary significantly if the Spatial analysis of archaeological data at the John Martin Dam and Reservoir, southeastern Colorado. texture of the parent material changes. Fur- Plains Anthropologist 29, 25-40. thermore, episodes of aggradation and degra- Fenneman, N.M. (1931). Physiography of Western dation in tributaries and other drainages can United States. New York: McGraw-Hill, 534 pp. Figgins, J.D. (1933). A further contribution to the an- be out-of-phase with those ofthe middle South tiquity of man in America. Proceedings Colorado Platte. Museum of Natural History 12, 4- 10. The middle South Platte River is an ideal Frison, G.C. (1978). Prehistoric Hunters of the High Plains. New York: Academic Press, 457 pp. setting within which to combine studies of Frison, G.C., and Stanford, D.J. (1982). Summary and late Quaternary stratigraphy and geoarche- conclusions. In G.C. Frison and D.J. Stanford, Eds., ology. An understanding of the geomorphic The Agate Basin Site, pp. 361-370. New York: Aca- history of the South Platte has aided in recon- demic Press. Gardner, M.E. (1967).Quaternary and Engineering Ge- structing the landscape setting during Paleo- ology of the Orchard, Weldona, and Ft. Morgan indian occupations and interpreting poten- Quadrangles, Morgan County, Colorado. Doctoral tially redeposited cultural debris. Moreover, Dissertation, Colorado School of Mines, Golden. Guthrie, R.L., and Witty, J.E. (1982).New designations the archaeological data have aided signifi- for soil horizons and layers and the new Soil Survey cantly in establishing the geochronology of Manual. Soil Science Society of America Journal 46, the river and the landscape evolution of the 443 -444. Hall, C.L. (n.d.1. The Geology of the Jurgens Site Based area. on Preliminary Evidence. Unpublished manuscript, University of Colorado Museum, 46 pp. I thank a number of individuals who provided help with Haynes, C.V. (1974). Archaeological Geology of some this research. Joe Ben Wheat (University of Colorado, Selected Paleo-Indian Sites. In C.C. Black, Ed., His- Boulder) suggested the research topic and he and Frank tory and Prehistory of the Lubbock Lake Site. The Frazier (American Stratigraphic, Denver) freely shared Museum Journal 15, 133- 140.

328 VOL. 2, NO. 4 MIDDLE SOUTH PLATTE RIVER

Haynes, C.V. (1980). The Clovis culture. Canadian the United States, pp. 243 -254. Princeton: Princeton Journal of Anthropology 1, 115-121. University Press. Haynes, C.V., and Haas, H. (1974).Southern Methodist Smith, R.O., Schneider, P.A., and Petri, L.R. (1964). University radiocarbon date list I. Radiocarbon 16, Ground-water resources of the South Platte River 368 - 380. Basin in western Adams and southwestern Weld Hunt, C.B. (1954). Pleistocene and Recent deposits in Counties, Colorado. U.S. Geological Survey Water the Denver area, Colorado. US.Geological Survey Supply Paper 1658, 132 pp. Bulletin 996-C, pp. 91- 140. Soil Survey Staff (1975). Soil Taxonomy. US. Depart- Hunt, C.B. (1974).Natural Regions of the United States ment of Agriculture, Soil Conservation Service, Agri- and Canada. San Francisco: W.H. Freeman and culture Handbook 436, 754 pp. Company, 725 pp. Spikard, L. (1972).Progress report of a Dent site investi- Machette, M.N. (1975a).The Quaternary Geology of the gation. unpublished manuscript, University of Colo- Lafayette Quadrangle, Colorado. Master’s Thesis, rado Museum, Boulder. University of Colorado, Boulder. Stanford, D.J. (1978).The Jones-Miller site: an example Machette, M.N. (1975b). Geologic map of the Lafayette of Hell Gap Bison procurement strategy. Plains An- Quadrangle, Adams, Boulder, and Jefferson Coun- thropologist Memoir 14, 90-97. ties, Colorado. US.Geological Survey Map MF-656. Van Horn, R. (1976).Geology ofthe Golden Quadrangle, Malde, H.E. (1954). Geological Notes on the Dent Site. Colorado. U.S. Geological Survey Professional Paper Unpublished data, University of Colorado Museum. 872, 116 pp. Malde, H.E. (1984).Geology of the Frazier site, Kersey, Wheat, J.B. (1972). The Olsen-Chubbock site: a Paleo- Colorado. In A.B. Anderson, Ed., Paleo-indian Sites Indian Bison kill. Society for American Archaeology, of the ColoradoPiedmont, pp. 13 - 16. Boulder: Amer- Memoir 26, 180 pp. ican Quaternary Association Field Trip Guidebook. Wheat, J.B. (1979).The Jurgens Site. Plains Anthropol- Muhs, D.R. (1985).Age and paleoclimatic significance of ogist Memoir 15, 153 pp. Holocene sand dunes in northeastern Colorado. An- Wilmsen, E.N., and Roberts, F.H.H. (1978). Linden- nals of the Association of American Geographers 75, meier, 1934- 1974: Concluding report of investiga- 566- 582. tions. Smithsonian Contributions to Anthropology 24, Roberts, F.H.H. (1937). New developments in the prob- 1- 187. lem of the Folsom Complex. Smithsonian Institution Wormington, H.M. (1957). Ancient Man in North Explorations and Field Work in 1936, pp. 67-74. America. Denver Museum of Natural History, Popu- Schumm, S.A. (1973). Geomorphic thresholds and com- lar Series 4, 322 pp. plex response of drainage systems. In M. Morisawa, Wormington, H.M. (1984).The Frazier site, Colorado, in Ed., Fluvial Geomorphology,, pp. 299-310. Bing- A.B. Anderson, Ed., Paleo-Indian sites of the Colo- hamton: State University of New York, Publica- rado Piedmont, pp. 12-13. Boulder: American Qua- tions in Geomorphology. ternary Association Field Trip Guidebook. Scott, G.R. (1963).Quaternary geology and geomorphic history of the Kassler Quadrangle, Colorado. US. Geological Survey Professional Paper 421 -A, 70 pp. Scott, G.R. (1965).Nonglacial Quaternary geology of the Southern and Middle Rocky Mountains. In H.E. Received April 3, 1987 Wright, Jr. and D.G. Frey, Eds., The Quaternary of Accepted for publication June 1, 1987

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 329